AI in Practice – Process Improvement to a 99.98% Accurate AI Model

AI in Practice – Process Improvement to a 99.98% Accurate AI Model

This bespoke AI model we created for an agricultural firm can detect plant diseases, pests and water levels by aggregating data from sensors and high-resolution cameras. On a valuation data set of 11k images, it is 99.98% accurate. Here’s how:

The Challenge

The challenges facing modern agriculture include water management, spotting and curing disease and pests, and labour shortages.

Farms can use an infra-red satellite image to spot land and crops which may be suffering drought, over-watering, pest infestation or disease. But specific detail is hard to achieve without sending someone out to look. Even then, growers can miss problems. The idea for the company was born out of a missed root fungus on a farm in Las Vegas, which resulted in the loss of a $17m crop.

The challenge for CKX was to work out an AI model, combined with robots, which could spot pests and diseases and monitor water levels. Farmworkers could then make one visit to the afflicted area to remedy the problem.

AI in Practice

To put the idea into practice, CKX and the agricultural firm developed a database of historical data and labelled 11K images of pests and diseases.

This knowledge is now combined with drones and ground-based robots. These machines can take the information to the afflicted area and asses the problem. The details are then sent in real-time to farmworkers who can deal with the problem. It’s one trip with the right kit.

The AI-guided robots can also asses maturity levels of plants and predict yields. This makes it easier for the farmer to price the harvest.

The robots can also carry portable DNA analysers, which is particularly important in vineyards. When machines and humans can’t work out what’s wrong, AI in the cloud replaces the distant lab. The robot can collect a soil or plant example, and the information can be checked, giving farmers real-time analysis, rather than having to wait critical weeks for the lab results to come back.

The CKX Process

For this solution, we combined off-the-shelf technology and a bespoke model. A user-friendly dashboard was also built using serverless technology that was hosted on Amazon Web Services (AWS). CKX is an AI and Amazon Web Services specialists. This gives clients:

? Significant cost reductions – save 90% on AWS bills

? Drastically reduced exposure to traffic spikes

? True scalability

? Top-level cybersecurity as standard with AWS compliance certification

We think one of the solutions to the challenges that modern agriculture faces can be a combination of AI and robotics. It’s AI as a force for good!

要查看或添加评论,请登录

Chris Knight的更多文章

社区洞察

其他会员也浏览了